Download Electromagnetic Actuators – Current Developments and examples

Electromagnetic Actuators – Current Developments and
examples
K. Hameyer1, M. Nienhaus 2
1
2
Katholieke Universiteit Leuven, EE dept., ESAT/ELECTA, Leuven-Heverlee, Belgium
mymotors & actuators gmbh, Wendelsheim, Germany
Abstract:
The range of applications for actuators covers all technical areas. Actuators can be found as system
component in nearly every controlled system to move or steer particular parts to control the overall
process of the system. Such actuators operate on various basic principles. However, in this paper only
actuators operated by the electromagnetic field are considered. Such devices are for example rotating
electrical machines, lifting magnets, active magnetic bearings for high-speed rotary applications,
magnetic levitation devices for traction applications and many more. Although large electromagnetic
devices, such as a large motor fits into this definition of an actuator, in this paper however only small
electromagnetic devices are discussed. It is intended to give a general overview over the development of
such small actuators. Models and methods to determine the operational behaviour of such devices are
discussed. Such computer models enable an improvement of the actuator’s function towards a desired
behaviour, objective to technical and/or economical constraints. Real applications, such as the PennyMotor as an ultra flat electromagnetic micromotor, illustrate the approach and demonstrate the general
and large range of applications for electromechanical actuators.
1000
3
Introduction
The ongoing trend towards the miniaturisation of
systems and the trend of system integration
force the development engineers to construct
smaller and smaller devices. Due to the recent
developments and ongoing research in the field
of hard- and soft-magnetic material delivered
materials with improved material properties
(Fig. 1). Therefore, extremely small electric
motors with very compact electromagnetic circuits can be developed (Fig. 3, 7, 8). Simultaneously developments in the area of the
system technology and their processes enable
actuator constructions at the on-chip level.
Micro-pumps can perform the crucial task of onchip cooling of processors operated by high
clock-frequencies. Devices with geometrical
dimensions in the range of some micrometer
based on the principles of the electrostatic field
can already be found on the market and have
gained technical reasonable maturity. Such
micro electro-mechanical systems (MEMS) are
applied e.g. as RF switches operate in GSM
applications, MEMS relays (Fig. 2) etc.
In contrast to such electrostatic applications in
the geometrical micrometer range, devices
operated by the electromagnetic field are larger.
A theoretical estimation of the favoured torque
ranges of the mentioned electrostatic and
kJ / m
900
(BH)max.theor. = 960 kJ / m³ theoretical
800
(BH)max. 75 % = 720 kJ / m³ technically possible
700
600
500
(BH)max / 2
400
NdFeB
300
Sm2Co17
200
AlNiCo
100
SmCo5
Steel
Ferrit
0
1880 1900 1920 1940 1960 1980 2000 2020 2040 2060 2080
Year
Fig. 1. Development of permanet magnet
material. (source VACUUMSCHMELZE)
electromagnetic actuation principles draws this
line between both at 1µNm [9]. Actuators
working on the electromagnetic principle can be
found in outer dimensions in the range of some
millimetre. A reason for this is that for those
small dimensions the energy stored in the
electromagnetic field is significantly smaller
when compared to that of the electric field.
However, the voltage exciting the electric field is
large when a high value for the field energy has
contacts
a technical or non-technical process. Following
this definition actuators are the technical
elements, which adjust and control e.g. a flow of
energy. Their input is a power-less electrical
signal and their output an electrical power or
mechanical energy, which is commonly a torque
for motors or a thrust for translational
movements.
Magnetic models computing the forces
driving capacitor
capacitor
Fig. 2: MEMS relays (total length < 0,1 mm source:
ESAT/MICAS KULeuven).
to be realised to have large forces for actuation
and this can be one of the constraints in the
design.
Nowadays, applications of small electromagnetic
actuators can be found in various fields. Novel
types of such devices are usually implemented
in systems out of the innovative industries, such
as the automotive industry, computer periphery
device
manufacturer,
audio / video
device
production and the medical instrumentation
industry. For such devices in general particular
properties regarding the quality, life-time, safety
of use, radiated audible noise and reliability are
prerequisites. To ensure such requirements and
to simultaneously decrease the time to market,
computer models serve during the development
as tools for the design engineer. A second
reason to apply computer simulations is the
usually complicated geometry of the electromagnetic actuators. To exploit the entire
properties of the material used in the
construction of an actuator sufficiently accurate
material models are required. Regarding the
increasing frequency operating the machines to
receive high speeds [5], particular attention must
be paid on the correct and accurate material
models. With further miniaturisation of the
electromagnetic devices special modellisation
techniques are required to simulate the device on
the crystal level.
Electromagnetic actuators
Janocha [1] defines an actuator as the
connecting link between information processing
system components of a controlling system and
To basically generate a mechanical energy by an
electromagnetic actuator a force must be
produced by a magnetic field. Forces can be
observed e.g. on current carrying conductors
exposed to a magnetic field. Such a force is
called Lorenz-force. Other forces can be
generated by a varying reluctivity of the magnetic
circuit along its path of movement; they are
called reluctance forces. Both principles are
applied to actuator designs. The magnetic field
for the first mentioned principle can be excited
by permanent magnet material or by windings,
which are formed out of various coils [2, 3, 4].
Usually both, Lorenz- and reluctance force
effects are superposed. Therefore, the determination of the change of the active air gap along
the relative movement of the rotor and stator
geometry is important for an accurate force /
torque calculation. It is evident, that complicated
geometries require advanced numerical methods
to determine the air gap forces accurately.
For this purpose, three- and two-dimensional
numerical models to compute the magnetic field
of the actuator can be applied. The design of
¹
¶
·
¸
¹
¶ stator
· permanent magnet system
¸ rotor
¹ winding
Fig. 3: Magnetic field solution of a small
electromagnetic stepper motor. (software
Magnet 5.4)
electromagnetic actuators by using the finite
element method (FEM) is discussed in this
paper. Fig. 3 illustrates the solution of the
computation of the magnetic field of a small
stepper motor. To rotate the permanent magnet
system, which is constructed to the rotor, the
two stator-windings are excited by a particular
sequence of stator currents. The magnetic field
is plotted for the situation of an excited upper
winding. The colours indicate the level and the
flux density distribution.
Thermal models
Regarding the trend of minimisation thermal
aspects of a design have to be considered as
well. Actuators are very often installed very close
to sources of thermal heat or are exposed to
high ambient temperatures. The automotive
industry e.g. demands for actuators, which can
operate in a range of – 40º up to + 180º C of the
ambient temperature without losing their performance. Due to the temperature dependent
properties of the materials, e.g. resistance of the
winding material or the magnetic properties of
the permanent magnet material used in the
devices, the physical behaviour of the actuator is
varying as well.
To form the entire model of the actuator, next to
the copper losses, the magnetisation and eddy
current losses in the ferromagnetic material have
to be considered to represent further heat
sources for the thermal model. Regarding the
demanded high rotational speeds and therefore
high frequencies operating the motors, those
losses become very important.
use of finite element models in three or two
dimensions is appropriate as well. With the help
of a thermal model of the device under study the
appropriate cooling conditions can be evaluated
(Fig. 4). The coupling of both, magnetic and
thermal FEM model is possible and yields a
numerical iteration to obtain the desired solution
[6, 7].
Electric circuit models
Typically, power electronic circuits operate
actuators. Regarding the time, these circuits
usually do not deliver a constant energy flow to
the device. Furthermore, actuators must not
operate continuously. Therefore, duty cycles
determine the thermal equilibrium of the device
and have to be considered in the overall model.
The behaviour of the actuator depends on the
power electronic as well. In the case induced
currents or voltages have to be computed in twodimensional FEM magnetic models external
winding circuits have to be defined. Such circuits
couple the FEM model to a linear network of
reactances, capacities, voltage sources and
resistances [8]. In these circuits the properties of
the power electronic components can be
considered.
Finite element models
To build up a FEM model, in a pre-process the
device’s geometrical shape is approximated by
finite elements. In two-dimensional models the
standard elements are commonly triangles and
in the three-dimensional case tetrahedron
elements can be used. This type of elements is
very flexible and capable to approximate very
well even complicated geometrical shapes
(Fig. 5). In this step of modelling the device,
material and boundary conditions have to be
defined.
Flux
Coil
Movement
Fig. 4: Detail of the thermal field solution of a
heat flow in a slot of an electrical machine. (inhouse software Olympos2D)
Due to system integration heat flows dissipated
from the power electronic components, which
might be fixed to the housing of the actuator
represent further heat sources in the thermal
model. For the modelling of thermal fields the
10
Fig. 5: (top) Geometry, direction of movement,
expected flux path and (below) the associated
3D FEM discretisation of a linear transversal flux
(TF) motor. The surrounding discretised air is not
plotted. (software Magnet5.4)
Net force (in N)
5
0
-5
-10
0
On the finite element a function is applied to
represent the desired potential describing
discretised Maxwell equations. This system of
equations has to be solved in a process to be
evaluated during the last step of analysis, the
post-process. Outgoing from the computed
potential solution, the interesting field values,
such as field strength and magnetic flux density
can be calculated for all places inside the model.
From the obtained field values e.g. the forces
and other physical quantities, such as the
induced currents and voltages can be derived.
Computer models enable a thorough study of
actuators without the requirement of expensive
prototyping. Prototyping is of course required at
the end of the design procedure to verify
simulated results or if required, to improve the
device’s numerical model. Various different
geometrical variants of an actuator concept can
be studied and the particular properties can be
compared. Figure 6 shows another variant of a
concept of a linear TF actuator with its
associated force characteristic.
Fig. 7: Computed flux density distribution inside
a rotating TF machine. Some parts of the
machine are made invisible to show the inner
yoke of the motor. (software Magnet6)
0,2
0,4
0,6
0,8
1
Position (in # pole pitches)
Fig. 6: 3D FEM models of a variant of a TF
linear motor with the associated force vs.
position characteristic.
With the computer models the appropriate
choice of material can be controlled and studied.
For example, the utilisation of the materials used
can be evaluated and the geometry may be
varied by keeping an identical actuator behaviour.
Fig. 7 shows the flux density distribution inside a
rotating TF actuator.
Penny-Motors
Designed to provide an output torque of more
than 100 µNm, the high-end Penny-Motor as a
member of an ultra flat micromotor family has a
diameter of 12,8 mm and a height of 1,4 mm
(Fig. 8). Important application fields for these
innovative drives are e.g. datacom, medicine,
automobiles, automation.
Fig. 8: Three members of the Penny-Motor
family: a) HE type for high torque; b) YL type for
low price; c) EC type for electronic commutation.
Besides the maximum torque and the ultra-flat
shape essential design parameters of the PennyMotors were a construction that allows easy
assembly and the suitability for massproduction.
This is the reason why all of the casing may be
produced by means of already industrially well
proven punching and bending techniques.
Core center of the penny motor is a three-strand
disk shaped coil produced by lithographical
methods and a merely 400 µm thick magnet-ring
made of rare-earth material consisting of eight
segments, which correspond to four poles-pairs.
Both coil and magnet-ring decisively determine
the performance of the motor and may be
optimally produced by a combination of precision
and microtechnical methods.
The integrated miniature ball bearing is as an
option magnetically pre-stressed and therefore
practically free of slackness. In addition, the
revolving magnetic back-iron minimizes the
influence of eddy currents with the known positive
effects on the motor’s efficiency (Fig. 9).
Magnetic circuit of the Penny-Motor
The main goal of dimensioning the magnetic core,
was to achieve a maximum torque without
rendering the assembly to costly. The operating
point of the Penny-Motor was calculated and
optimized by parameter variations. The calculation
includes an approximation of the stray materials,
such as rare-earth permanent magnet material
NdFeB (Fig. 1). Different types of soft magnetic
materials where studied as well.
Fig. 9: Magnetic circuit of the HE Penny-Motor.
To obtain a better knowledge on the magnetic field,
a commercial FEM tool was used to examine the
different motor types. Verified by experiments the
magnetic flux density within the air-gap of the highend motor shows signifi cantly more than 0,5 T. As
illustrated in Fig. 9, the soft magnetic components
lead the magnetic flux essentially through the
ferromagnetic housing of the motor. These soft -
Fig. 10: Function of the non rotating pre-stress ring
within the magnetic circuit of the Penny-Motor.
(software ANSYS)
magnetic parts of the motor are optimized for a
saturation of 1,6 T. This design ensures that the
bearing in the inner area is not affected by magnetic
fields. Fig. 10 shows the effect of the pre-stress
ring. In this configuration the ring causes a well
defined magnetic force of about 0,3 N on the ball
bearing. Depending on the geometry of the prestress ring this force can be varied in the wide
range of 0 N up to more than 5 N.
Magnetizing miniaturized NdFeB magnets
An innovative method was applied to produce the
required NdFeB magnet-rings including the
magnetization procedure is to employ enhanced
micro moulding techniques and an already
patented magnetizing method as described in [10,
11]. A disadvantage of the type of compound
magnets which could not be solved yet is the
relatively low energy product and the high magnetization energy required to entirely magnetize the
material when compared to the uncompounded
anisotropic NdFeB-material.
In the case of applying the anisotropic NdFeBmaterial the magnet-rings should be constructed of
a single ring guaranteeing for its geometry of being
within close tolerances. In spite of the lower
magnetization energy, when compared to the
compounded grades of NdFeB, the anisotropic
material requires a minimum magnetic field of e.g.
2400 kA/m for 100% saturation. To ensure a
uniform, magnetization within the complete magnet
segment area, in practice magnetic fields up to
4000 kA/m and more were applied to the magnets
during the magnetization procedure.
Because of the small magnet segments and the
required high magnetic field, a specially adapted
magnetization tool was developed [11]. To save
time and costs a commercial FEM code (Fig. 11)
was employed to deliver simulation results of the
magnetic field excited by high currents. This
represented a solid basis for the final design of the
magnetization tool.
Fig. 11: 3D profile of a magnetic field caused by a
current of approximately 5000 A through a special
shaped copper conductor within the magnet ring.
(software MAXWELL3D)
Penny-Motor bearings
Various bearing configurations have been considered under technical, economical and under
availability aspects. For a ball bearing it can be
shown that deep grooved radial ball bearings in the
dimensions of the bore of 1 mm, outer diameter
3 mm, width 1 mm and the precision class of ISO
P4 / ABEC 7 represent a standard bearing with
features already close to the requirements for the
Penny-Motor. A critical feature of the Penny-Motor
assembly is the amount of angular misalignment
within the bearings. Referring to a ball bearing, the
angular misalignment is defined as the degree to
which the outer ring in a diagonal way can be
displaced versus the inner ring. For the PennyMotor this displacement results in a corres ponding
lateral displacement of the rotor disc. The
relationship can be determined by the tangent of the
lateral displacement d = r ⋅ tan β (Fig. 12).
alignment that can be expected in the Penny-Motor
bearings as a function of the radial slackness tolerance. Varying the radial slackness from 1 to 10 µm
and assuming unpreloaded bearings with a
standard curvature of 57% the angular misalignment lies within a range of approxi mately 0,5°
to 1,5°.
This seems not to be the optimum for the PennyMotors and that is the reason why also high quality
standard ball bearings are not matching the
motor’s requirements. Therefore, an extensive
development of the ball bearing focussing on the
Penny-Motor has been performed. The use of
tighter raceway curvatures can reduce the displacement by more than 50%. Other important optimization parameters are the tolerances of the inner
and outer ring diameter of the bearings and of the
corresponding motor com ponents. This tolerances
causes during the final assem bly complex and
significant deformations with an extensive influence
on the resulting radial slackness of the bearing
(Fig. 13).
Employing a high-end measuring
machine with an automated clas sification and
sorting routine allows the assignment of well suited
parts of the bearing system for the final assembly to
guarantee high quality bearings with a minimized
angular mis alignment.
Conclusion
Computer simulation is a powerful assistant in the
field of micro actuator R&D. Decisively for success
is often the combination of simulating electric,
electromagnetic, thermal and mechanic phenomena’s. It is not compelling to solve this tasks with
only one flexible but complex tool. In many cases of
practical relevance a combination of highly
specialized simulation tools leads to qualified
results.
Fig. 12: Angular misalignment of a ball bearing.
The magnitude of the angular misalignment is a
function of the internal geometry of the bearing,
radial slackness influenced by the assembly and
cross raceway curvatures. An analysis has been
performed to determine the degree of angular mis-
Fig. 13: Influence of mechanical assembly on the
internal geometry of a minature ball bearing.
(software ProMechanica)
Acknowledgements
The authors are grateful to the Belgian “Fonds voor
Wetenschappelijk Onderzoek Vlaanderen” (project G.0420.99) and
the Belgian Ministry of Scientific Research (project IUAP No.
P4/20). The authors thank ir. Hans Vande Sande and ir. Geoffrey
Deliege for the carefully performed computations for the examples
in the Fig.’s 3-6. The authors thank also Dr. J. Tenbrink and Dr. F.
Jurisch from Vacuumschmelze GmbH & Co. KG in Germany for
their input concerning soft and hard magnetic materials.
References
[1]
[2]
H. Janocha, Aktuatoren, Springer Verlag, 1992.
E. Hering, A. Vogt, K. Bressler, Handbuch der Elektrischen Anlagen und Maschinen, Springer Verlag
1999.
[3] H.-D. Stölting, E. Kallenbach, Handbuch Elektrische
Kleinantriebe, Hanser Verlag, 2001.
[4] W.H. Yeadon, A.W. Yeadon, Handbook of small
electric motors, McGraw-Hill, 2001.
[5] P.-A. Paratte, A. Birkicht, “Electromagnetic high
speed micromotor characterisation”, Proc. Conf.
MOVIC ’98, 4th Int. Conf. on motion and vibration
control, August 25-28, 1998, ETH Zürich,
Switzerland.
[6] Driesen J., Deliège G., Belmans R., Hameyer K.:
“Coupled thermo-magnetic simulation of a foil-winding
transformer connected to a non-linear load,” IEEE
Transactions on magnetics, vol.36, No.4, July 2000,
pp.1381-1385.
[7] K.Hameyer, J.Driesen, H.De Gersem, R.Belmans: “The
classification of coupled field problems,” IEEE
Transactions on Magnetics, Vol.35, No.3, May 1999,
pp.1618-1621.
[8] De Gersem H., Hameyer K.: “Electrodynamic finite
element model coupled to a magnetic equivalent
circuit,” The European Physical Journal, vol.12, No.2,
November 2000, pp.105-108.
[9] Hagemann, B.: Entwicklung von PermanentmagnetMikromotoren mit Luftspaltwicklung, Diss. Hannover,
1997.
[10] Nienhaus, M.; Ehrfeld, W.; Stölting, H.-D.; Michel, F.;
Kleen, S.; Hardt, S.; Schmitz, F.: “Design and Realization of a Penny-Shaped Micromotor”. Proc. SPIE
Vol. 3680B-65, Paris, France, 1999.
[11] Kleen, S.; Ehrfeld, W.; Michel, F.; Nienhaus, M.;
Stölting, H.-D.:“Penny-Motor: “A Family of Novel
Ultraflat Electromagnetic Micromotors”. Actuator 2000,
Bremen, 2000.